Volatile profiles of sparkling wines obtained by three extraction methods and gas chromatography mass spectrometry (GC MS) analysis

Food Chemistry 105 (2007) 428 435 Analytical, Nutritional and Clinical Methods Volatile profiles of sparkling wines obtained by three extraction methods and gas chromatography mass spectrometry (GC MS) analysis Joan Bosch-Fusté a, *, Montserrat Riu-Aumatell a, Josep M. Guadayol b, Josep Caixach c, Elvira López-Tamames a, Susana Buxaderas a a Dept. Nutrició i Bromatologia, Universitat de Barcelona, Avda. Joan XXIII, s/n, 08028 Barcelona, Spain b Dept. Enginyeria Química, Universitat Politècnica de Catalunya, Colom, 1, 08222 Terrassa, Spain c Laboratori d Espectrometria de Masses, IIQAB-CSIC, c/jordi Girona 18, 08034 Barcelona, Spain Received 30 June 2006; received in revised form 29 November 2006; accepted 27 December 2006 Food Chemistry www.elsevier.com/locate/foodchem Abstract Simultaneous distillation extraction (SDE) and closed-loop stripping analysis (CLSA) show great capacity for organic compound extraction. Here we used these techniques to obtain and characterize a wide range of volatile compounds from aged cava sparkling wine. We also explored the potential application of head space-solid phase micro extraction (HS-SPME) to determine the distinctive volatile compounds of aged cava as this technique is a common extraction method in quality flavour control. For SDE, 50 ml of cava were extracted with pentane/dichloromethane during 4 h in a Likens Nickerson (LN) extraction apparatus; while for CLSA, 25 ml of cava were stirred during 1 h into a CLSA apparatus with an adsorbent trap of granulated activated carbon. HS-SPME was performed at 35 C using 2 ml of sample for 30 min. The 20-mm SPME fibber was coated with a 50/30-lm layer of divinylbenzene carboxen polydimethylsiloxane (DVB CAR PDMS). All the extraction methods were followed by gas chromatography mass spectrometry (GC MS) analysis. SDE and CLSA allowed the identification of 84 volatile compounds. Almost 40% of the volatiles from these two techniques were obtained by HS-SPME. Moreover, here we provide the first description of several tentatively identified compounds such as lilial, octanal, 2-octanone, isopropyl disulfide, methylthiophen-3-one, a-amyl-cinnamaldehyde, ethyl 2-furancarboxylate, 2-acetylfuran, and 5-methylfurfural in cava. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Volatile compounds; HS-SPME; CLSA; SDE; Sparkling wine; GC MS 1. Introduction The volatile composition of wines is complex because several organic structures contribute to flavour (Aznar, López, Cacho, & Ferreira, 2001; Câmpeanu, Burcea, Doneanu, Nămolosanu, & Visan, 1998; Castro, Natera, Benitez, & Barroso, 2004; Genovese, Dimaggio, Lisanti, Piombino, & Moio, 2005; Karásek et al., 2003; Ortega, López, Cacho, & Ferreira, 2001; Salinas, Alonso, & Esteban-Infantes, * Corresponding author. Tel.: +34 93 402 45 08; fax: +34 93 403 59 31. E-mail address: jbosch001@ub.edu (J. Bosch-Fusté). 1994; Sánchez-Palomo, Pérez-Coello, Díaz-Maroto, González-Viñas, & Cabezuedo, 2006; Schneider, Baumes, Bayonove, & Razungles, 1998; Zea, Moyano, Moreno, Cortes, & Medina, 2001). The amounts of volatile compounds in wine range from nanograms to micrograms; moreover, these compounds have distinct physicochemical properties regarding, for example, polarity, volatility, and odour impact (Aznar et al., 2001; López, Ferreira, Hernández, & Cacho, 1999) as a result of the functional groups (alcohol, aldehyde, acid, etc.) present in the molecules. Cava is a quality sparkling wine (Certified Brand of Origin) elaborated by the traditional method that consists of a second fermentation followed by biological ageing in contact with lees in 0308-8146/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodchem.2006.12.053

J. Bosch-Fusté et al. / Food Chemistry 105 (2007) 428 435 429 anaerobic conditions for at least 9 months (Council Regulation (EC), 1493/1999). Volatile compounds in wine have three origins: from the grape (pre-fermentative aroma); from the yeast during the first or second fermentation (fermentative aroma); or from ageing during settling (post-fermentative aroma). This special ageing of cava gives this wine a more complex volatile profile since autolytic and enzymatic reactions may take place. It is thought that the base wine, the autolysis of lees of the second fermentation and the ageing time in contact with these lees are the most important factors that affect the sensory quality of sparkling wines. Aroma is also of paramount importance for producers of high quality sparkling wines such as champagne (Vannier, Brun, & Feinberg, 1999). In this regard, the flavour of cava is empirically described by cava makers with yeasty and toasty notes. Although several volatiles, like 1,2-dihydro- 1,1,6-trimethylnaphthalene (TDN), vitispiranes isomers and diethyl succinate, have been used as markers of ageing (Riu-Aumatell, Bosch-Fusté, López-Tamames, & Buxaderas, 2006), distinctive compounds involved in the flavour of aged cava have not been characterized to date. Here we applied a simultaneous distillation extraction method (SDE) and closed-loop stripping analysis (CLSA), both with high uptake techniques, to study a wide volatile profile of aged sparkling wine. SDE (Likens Nickerson technique) combines the advantages of liquid liquid and steam distillation extraction, and has been used to isolate volatile organic compounds from spices (Díaz-Maroto, Pérez-Coello, & Cabezudo, 2002) and wine (Blanch, Reglero, & Herraiz, 1996), with very high recovery rates. CLSA has been used to extract pollutants at very low concentrations from water samples (Malleret, Bruchet, & Hennion, 2001). Recently CLSA has been proposed as an effective method for isolating wine aroma compounds (Eggers, Kenefick, Richardson, Wigglesworth, & Girad, 2003). However, neither SDE nor CLSA have been applied to obtain a volatile profile from cava sparkling wines. Head space-solid phase micro extraction (HS-SPME) is a simple, rapid, solvent-free and inexpensive method for extracting volatile and semi-volatile compounds. It is based on equilibrium between the analyte in vapour phase and silica fibre coated with an adsorbent polar or apolar polymer. A wide range of commercial fibres are available on the market, some extract the target analytes by partitioning while others extract the volatile compounds by physical trapping and also by partitioning. Finally, the analytes are directly desorbed in the injector port of GC apparatus. HS-SPME is the most common and easiest technique for food quality control (Liu, Zeng, & Tian, 2004; Tat, Comuzzo, Stolfo, & Battistutta, 2005; Torrens, Riu-Aumatell, López-Tamames, & Buxareras, 2004) and in the last few years, this method has been applied to study the compounds involved in wine flavour (Bonino et al., 2003; De la Calle-García et al., 1997; Vas, Gál, Harangi, Dobó, & Vékey, 1998). Here we analysed aroma extracts of aged cavas by means of SDE and CLSA techniques in order to identify new compounds. These extraction techniques were applied to the same cavas in order to later compare the results with qualitative data obtained by HS-SPME. 2. Materials and methods 2.1. Samples We used a total of three cavas (Spanish sparkling wine with an ageing time of 14 months), manufactured by the same winery on an industrial scale from the autochthonous Vitis vinifera of the cava region. Cavas were made with a blend of the traditional white varieties: Macabeu, Xarello and Parellada (1:1:1). Amber glass bottles of 250 ml were filled with 200 ml of cava and the headspace was saturated with carbon dioxide. The bottles were stored at 37 C in a heater and samples were removed at 0 days (control cava) and after 30 days (aged cava) and stored at 20 C until analysis. All glass bottles with samples were spiked with 2.5 mg L 1 of 2-octanol (98% purity) as internal standard (IS). 2.2. Chemicals and reagents Ethyl propionate, ethyl butyrate, propanol, isobutyl alcohol, hexanal, isoamyl acetate, methyl hexanoate, butanol, limonene, isoamyl alcohol, ethyl hexanoate, pentanol, hexyl acetate, octanal, ethyl lactate, hexanol, trans-3-hexenol, cis-3-hexenol, methyl octanoate, ethyl octanoate, heptanol, furfural, ethyl nonanoate, octanol, c-butyrolactone, 5-methylfurfural, methyl decanoate, ethyl decanoate, isoamyl octanoate, diethyl succinate, a-terpineol, undecanal, decanol, 2-phenylethyl acetate, ethyl dodecanoate, benzyl alcohol, hexanoic acid 2-phenylethanol, methyl tetradecanoate, octanoic acid, methyl hexadecanoate, decanoic acid, methyl octadecenoate, tetradecanoic acid, hexadecanoic acid, and 2-octanol (IS) were purchased from Sigma Aldrich and Fluka (St. Louis, MO, USA) and showed purity higher than 95%. Individual stock standard solutions of each aroma compound were prepared by weight in methanol (SDS, Peypin, France). A commercial solution of 24 aliphatic hydrocarbons (C8 C32) in hexane (Supelco, Bellefonte, PA, USA) was added to the samples in order to calculate Kovat s indices. 2.3. Extraction methods SDE: 50 ml of cava were placed in a 250-mL flask with 100 ml of distilled water. A second flask with 5 ml of a mixture of pentane and dichloromethane (3:1) (SDS, Peypin, France) was also attached to a Likens Nickerson apparatus. Solvent and sample were heated to their boiling points. These temperature conditions were maintained for 4 h. A solution of polyethylenglycol in water was used as cooler to condensate vapours and to avoid loss of volatile compounds. The extract was then collected at room temperature, dried over sodium sulphate anhydrous (Panreac S.A., Montcada i Reixac, Spain) and concentrated up to 0.5 ml under nitrogen gas.

430 J. Bosch-Fusté et al. / Food Chemistry 105 (2007) 428 435 CLSA: Extraction was performed in a commercial CLSA apparatus (Brechbüler, Zurich, Switzerland). Twenty five ml of cava were diluted up to 1000 ml with double distilled water. The samples were air-stirred (flow rate of 1.5 L/min) for 70 min in a bath at 45 C. Conditions were similar to those used by Eggers et al. (2003) in wine model solution, although in our case the adsorbent trap was a 5-mg filter of activated charcoal at 55 C.Thefilterwasthenextracted with 40 ll of carbon disulphide (SDS, Peypin, France). The wine was diluted 1/40, so the ethanol concentration was not enough to alter the solubility of analytes. In these conditions non-saturation of activated charcoal was expected. HS-SPME: Conditions were similar to those used by Riu-Aumatell et al. (2006). Two millilitres of sparkling wine were placed into a 10 ml glass vial (Reference 27385, Supelco, Bellefonte, PA, USA) for each HS-SPME analysis. A small magnetic stirring bar was also added. The vial was tightly capped with a PTFE septum and placed in a water bath with stirrer. The sample was maintained for 30 min at 35 C. The fibre was activated by inserting it into the GC injector at 280 C for 30 min. The SPME fibber was then inserted into the headspace. During the sampling time (30 min), the cava sample was stirred at constant speed (700g). After reaching the sampling time, the fibre was removed from the vial and inserted into the GC injection port for desorption during 5 min. A 20 mm-fibre coated with a 50/30 lm layer of divinylbenzene carboxen polydimethylsiloxane (DVB CAR PDMS) was used. 2.4. GC MS analysis Volatiles were identified on a mass spectrometer Agilent Technologies 5973 Network coupled directly online to an Agilent Technologies 6890N Network GC System (both, Palo Alto, CA, USA). Spectra were obtained on electron impact at 70 ev, scanning from 15 to 250 m/z at 2 scans s 1. The GC system was equipped with a supelcowax 10 (Supelco, Bellefonte, PA, USA) capillary column with a 20 M polyethyleneglycol stationary phase (30 m 0.25 mm 0.25 lm) and SPB-1 (Supelco, Bellefonte, PA, USA) fused silica capillary column (30 m 0.25 mm 0.25 lm). The temperature programme ranged from 40 C to 250 C in the following way: 40 C for 10 min, from 40 C to 200 C at2 C min 1, one min at 200 C, from 200 C to 250 C at2 C min 1, and finally, 10 min at 250 C. The carrier gas was helium at 1 ml min 1. One microlitre of each extract from SDE and CLSA was injected in splitless mode. An 8-min solvent delay was programmed when the SDE extracts were injected. 2.5. Identification Volatiles were identified by comparing the retention index and mass spectrum of commercial standards when they were available. The volatile compounds were also identified using the software library of mass spectra database Willey 6.1 (NY, USA); moreover, they were identified using theoretical retention index calculation on polar column (Supelcowax-10 described above) and on non-polar column (SPB-1 described above). In addition, another GC MS integrated quadrupole Trace MS Plus apparatus (ThermoElectron, USA) was used to confirm the identification of the volatile compounds. The following temperature programme was applied: 35 C for 5 min, up to 250 C at 5 C min 1, and finally, 20 min at 250 C. A DB-5 column (60 m 0.25 mm 0.25 lm) (J & W scientific, Folsom, CA, USA) was used with this GC MS system. The MS parameters were: a mass range of 35 350 m/z, with 0.5 scan s 1, and ionization energy of 70 ev. NIST/EPA/ NIH mass spectral database 1998 (Gaitherfburg, MD, USA) library was used with this GC MS equipment. 3. Results and discussion Three extraction methods (SDE, CLSA, and HS-SPME) were used to identify volatile compounds in cava (Table 1). The compounds are listed following elution order, and including their chemical name, CAS number (SciFinder Scholar Ò 2006) or mass fragments, Kovat s index for polar and non-polar columns, cited Kovat s and the identification method used. Chromatograms from SDE, CLSA, and HS-SPME are shown in Fig. 1. The peak numbers of the chromatograms match the numbers of Table 1. The semi-quantitative analysis was performed by internal normalization and is shown in Table 1 by means of an asterisk scale (<1%; 1 10%, and >10%). SDE required 50 ml of sample for extraction while CLSA and SPME needed 25 and 2 ml, respectively. Therefore, SDE allowed us to obtain a cava chromatogram with a wider volatile profile than those from CLSA and SPME (Fig. 1). However, SDE is a lengthy process (4 h); moreover, this method requires solvents and temperature for volatile compounds extraction. In this study we used dichloromethane because of its polarity. Consequently, high extraction capacity for polar volatile compounds (alcohols, aldehydes, and acids) was obtained. This capacity was not observed in CLSA and SPME because these methods were based on the equilibrium partitioning of the volatiles between solution phase (sparkling wine) and vapour or gas phase. This equilibrium is regulated by the partition coefficient and thus it is difficult to apply HS- SPME and CLSA for the extraction of polar compounds from a hydro-alcoholic matrix like cava. CLSA (stripping technique) and HS-SPME identified only the most abundant alcohols such as propanol, isopropanol, isoamyl alcohol, hexanol, cis-3-hexenol and 2-phenylethanol. In contrast, in SDE extracts some ramified alcohols, like 2- methyl-1-pentanol, 3-methyl-1-pentanol, and isohexanol (4-methyl-1-pentanol) were detected (Table 1). These alcohols have been described in wine (Tat et al., 2005), sherry (Zea et al., 2001), and cognac (Ledauphin et al., 2005), and their occurrence could be related to oxidative ageing (Zea et al., 2001). Moreover, tentatively identified 3-ethoxy-1-propanol in the SDE extracts has been reported in

Table 1 Volatile compounds identified by SDE, CLSA, and HS-SPME Cas No./MS A IK CW B Cited IK C IK SPB1 D ID E Control cava Aged cava SDE CLSA SPME SDE CLSA SPME 1 Ethyl propanoate 105-37-3 <1000 f, g * * * * 2 Ethyl isobutyrate 97-62-1 <1000 965 F g, h * * * * 3 Butyraldehyde, diethyl acetal 3658-95-5 <1000 1031 G g, h * * 4 Ethyl butyrate 105-54-4 1031 1052 H 800 f, g, h * * * * ** * 5 Propanol 71-23-8 1046 1069 H f, g, h ** * ** * 6 Ethyl 2-methylbutanoate 7452-79-1 1055 1052 F 820 g, h * * * * 7 Ethyl isovalerate 108-64-5 1061 1070 F 829 g, h * * * * * 8 Isobutyl alcohol 78-83-1 1108 1099 H f, g, h ** * * ** * * 9 Valeraldehyde, diethyl acetal 3658-79-5 1120 1135 G g, h * * 10 Isovaleraldehyde, diethyl acetal 03842-03-3 1079 1062c g, h * * 11 Hexanal 66-25-1 1094 1064 G f, g, h * 12 Isoamyl acetate 123-92-2 1123 1142 H 851 f, g, h * ** * * ** * 13 Methyl hexanoate 106-70-7 1101 1176 G 906 f, g, h * * * 14 Butanol 71-36-3 1159 1147 G f, g, h * * 15 1-Pentene-3-ol 918-85-4 1174 1157 I g, h * * 16 Limonene 138-86-3 1193 1175 G 1030 f, g, h * 17 Isoamyl alcohol 123-51-3 1210 1211 J 801 f, g, h *** *** ** *** *** ** 18 Ethyl hexanoate 123-66-0 1232 1239 H 977 f, g, h * *** *** * *** *** 19 Isopropyl disulfide 4253-89-8 1248 g * 20 Hexanal, diethyl acetal 3658-93-3 1250 1235 G g, h * 21 Pentanol 71-41-0 1258 1249 G f, g, h * 22 Hexyl acetate 142-92-7 1269 1267 J 1012 f, g, h ** * * * 23 2-Octanone 111-13-7 1278 1278 I 962 g, h * * * * ** * 24 Ethyl 3-hexenoate 2396-83-0 1286 1291 G g, h * * * 25 Octanal 124-13-0 1290 1282 G 973 f, g, h * 26 Isohexanol 626-89-1 1317 1312 G 838 g, h * * 27 2-Penten-1-ol 20273-24-9 1321 g * 28 1-Pentanol, 3-methyl- 589-35-5 1329 1323 G 852 g, h * * * * 29 Heptanal, diethyl acetal 688-82-4 1334 1332 G g, h * 30 Ethyl 2-hexenoate 1552-67-6 1336 g * * 31 2-Methyl-1-pentanol 105-30-6 1337 837 g * 32 Ethyl lactate 97-64-3 1348 1367 H 821 f, g, h ** * * ** * 33 Hexanol 111-27-3 1351 1371 H 845 f, g, h * ** * ** ** * 34 trans 3-Hexenol 928-97-2 1367 1385 H 854 f, g, h * * 35 3-Ethoxy-1-propanol 111-35-3 1375 840 g * * 36 cis-3-hexenol 928-96-1 1379 1401 J 853 f, g, h * * * * * * 37 Methyl octanoate 111-11-5 1383 1386 G 1103 f, g, h * * * * * 38 Ethyl octanoate 106-32-1 1430 1436 J 1178 f, g, h ** *** *** * *** *** 39 Heptanol 111-70-6 1458 1454 G 957 f, g, h * * * * * 40 Furfural 98-01-1 1459 1474 F 828 f, g, h * * * * * * 41 2-Ethylhexen-1-ol 29594-61-4 1478 g * * 42 2-Acetylfuran 1192-62-7 1490 1500 G 892 g, h * 43 Vitispirane 1 65416-59-3 1508 1507 J 1248 g, h * * * * * * 44 Vitispirane 2 65416-59-3 1511 1510 G 1252 g, h * * * * * * 45 Methylthiophen-3-one 13679-85-1 1519 1538 G 937 g, h * * 46 Nonanal, diethyl acetal 54815-13-3 1530 g * 47 Ethyl propionate 123-29-5 1536 1530 G f, g, h * (continued on next page) J. Bosch-Fusté et al. / Food Chemistry 105 (2007) 428 435 431

Table 1 (continued) Cas No./MS A IK CW B Cited IK C IK SPB1 D ID E Control cava Aged cava SDE CLSA SPME SDE CLSA SPME 48 1-Octanol 111-87-5 1561 1559 G 870 f, g, h * * * * * * 49 c-butyrolactone 96-48-0 1568 1634 I 835 f, g, h * 50 5-Methylfurfural 620-02-0 1572 1566 G 926 f, g, h * * 51 Diethyl malonate 105-53-3 1582 1043 g * * 52 Methyl decanoate 110-42-9 1593 1586 G 1302 f, g, h * * * * * 53 Ethyl 2-furancarboxylate 614-99-3 1621 1621 G 1009 g, h * * * * 54 Ethyl decanoate 110-38-3 1639 1678 H 1376 f, g, h * ** ** * * ** 55 Isoamyl octanoate 2035-99-6 1652 1649 J 1429 f, g, h * * 56 Diethyl succinate 123-25-1 1678 1662 J 1146 f, g, h ** ** * ** ** * 57 Ethyl 9-decenoate 67233-91-4 1691 1681 J 1360 g, h * * ** * * * 58 Terpene compound (ms) 93, 109, 204 1699 g * * 59 a-terpineol 7785-53-7 1710 1694 G 1185 f, g, h * 60 Terpene compound (ms) 105, 161, 204 1717 g * 61 Terpene compound (ms) 93, 161, 204 1723 1468 g * 62 TDN 30364-38-6 1731 1719 J 1313 g, h * * * * 63 Undecanal 112-44-7 1747 g * 64 Terpene compound (ms) 119, 132, 202 1756 1456 g * * 65 1-Decanol 112-30-1 1765 1764 G 1263 f, g, h * * * 66 2-Phenylethyl acetate 93-92-5 1811 1795 J 1222 f, g, h * * * * 67 Ethyl dodecanoate 106-33-2 1827 1833 J 1576 f, g, h * * 68 Benzyl alcohol 100-51-6 1870 1881 G 1033 f, g, h * 69 Hexanoic acid 142-62-1 1885 1826 J 1007 f, g, h ** * ** * 70 2-Phenylethanol 60-12-8 1906 1888 J 1074 f, g, h ** ** * ** ** ** 71 Isopropyl myristate 110-27-0 1936 1964 H g, h * * 72 Methyl tetradecanoate 124-10-7 2009 1682 f, g * 73 Lilial 80-54-6 2037 g * 74 Ethyl tetradecanoate 124-06-1 2049 2046 G 1779 f, g, h * 75 Octanoic acid 124-07-2 2087 2038 J 1164 f, g, h ** ** ** * * ** 76 p-vinylguaiacol 7786-61-0 2192 2200 G 1274 * 77 Methyl hexadecanoate 112-39-0 2216 2213 G 1909 f, g, h * * 78 Terpene compound (ms) 119, 161, 204 2234 g * * 79 a-amyl-cinnamaldehyde 122-40-7 2247 1598 g * * 80 Ethyl hexadecanoate 628-97-7 2255 2229 J 1979 f, g, h * * 81 Decanoic acid 334-48-5 2339 2254 J 1349 f, g, h ** *** ** * 82 Methyl octadecenoate 27234-05-5 2390 2417 G f, g, h * 83 Tetradecanoic acid 544-63-8 >2600 >2600 G 1731 f, g, h * * 84 Hexadecanoic acid 57-10-3 >2600 1919 g * * 432 J. Bosch-Fusté et al. / Food Chemistry 105 (2007) 428 435 The semi-quantitative analysis was performed by internal normalization and it was expressed by means of an asterisk scale (<1%; 1 10%, and >10%). f: identification based on retention time and electron impact mass spectrum of standards; g: identification based on examination of electron impact mass spectrum; h: identification based on theoretical retention index calculation. A CAS number or mass fragments of identified compounds. B Kovat s indices on Supelcowax-10 column. C Kovat s indices of polar columns reported in literature. D Kovat s indices on SPB-1 column. E Identification method. F Aznar et al. (2001). G Ledauphin et al. (2005). H Câmpeanu et al. (1998). I Comuzzo et al. (2006). J Riu-Aumatell et al. (2006).

J. Bosch-Fusté et al. / Food Chemistry 105 (2007) 428 435 433 a 2e+5 u.a. 5 8 17 18 32 IS 33 38 56 70 75 SDE 50 ml of sample 83 1.5e+5 SD 40 54 69 81 1e+5 5e+4 14 10 12 11 4 9 36 46 20 49 29 37 25 45 23 27 34 44 48 39 53 21 28 43 42 51 63 57 62 64 74 69 72 77 79 76 10 20 30 40 50 60 70 80 90 100 66 68 84 Time b 2e+5 u.a. 12 17 18 33 IS 38 54 56 81 CLSA 25 ml of sample 1.5e+5 c 1e+5 5e+4 2e+5 1 2 u.a. 4 7 6 8 13 10 20 30 30 40 50 60 70 80 90 100 17 23 32 37 22 36 24 28 18 IS 48 39 43 52 41 44 53 40 38 55 54 57 65 66 67 70 69 71 75 Time SPME 2 ml of sample 1.5e+5 57 1e+5 5e+4 1 2 4 5 6 78 12 33 37 22 23 32 36 52 56 40 43 44 48 55 62 65 67 10 20 30 40 50 60 70 80 90 Fig. 1. Chromatograms of sparkling wines: (a) chromatogram of a sparkling wine extract obtained by SDE; (b) chromatogram of a sparkling wine extract obtained by CLSA; (c) chromatogram of a sparkling wine obtained by HS-SPME. Peaks numbers are identified according to Table 1. 70 69 75 100 Time wine that has undergone malolactic fermentation (Fernandes, Relva, Gomes da Silva, & Costa Freitas, 2003). Finally, two sulphur compounds (isopropyl disulfide and methylthiophen-3-one) were detected in the SDE extracts. The presence of methionine and yeast metabolism could explain the formation of these sulphur compounds (Bayonove, Baumes, Crouzet, & Günata, 2000; Blaise & Bertrand, 2000). The highest amounts of polar compounds were obtained by SDE; however, HS-SPME showed the most representative polar compounds of cava. Indeed, it may be a useful extraction method for alcohols and acids. However, the high extraction capacity of SDE and CLSA for acid compounds (hexanoic, octanoic and decanoic acids) may mask the detection of other minority compounds as a result of peak tailing (Fig. 1, peaks 69, 75, 81). Ethyl esters of aliphatic acids are an important aromatic family in the volatile profile of sparkling wines. Although SDE, CLSA, and HS-SPME were suitable methods for determining ethyl esters (ethyl hexanoate, ethyl octanoate, and ethyl decanoate) in cava samples, SPME was the fastest and easiest technique to extract the main esters of volatiles. Other ethyl esters detected using these three extraction methods were ethyl 9-decenoate, ethyl lactate and diethyl succinate. The first ethyl ester has been reported in sparkling wine samples (Riu-Aumatell et al., 2006), and in white wines (Câmpeanu et al., 1998). Ethyl lactate and diethyl succinate are post-fermentative volatiles formed during the ageing of cava in contact with lees from the second fermentation. The concentration of diethyl succinate increases during ageing and this compound is a

434 J. Bosch-Fusté et al. / Food Chemistry 105 (2007) 428 435 marker of the evolution of cava during cellar storage (Francioli, Guerra, Lopez-Tamames, Guadayol, & Caixach, 1999; Riu-Aumatell et al., 2006). Other ester compounds present in the volatile profile of cava were methyl esters of hexanoic, octanoic, and decanoic acids. Ledauphin et al. (2005) detected methyl esters in spirit distilled beverages. In these beverages, the formation of methyl esters is due to the high temperature required during the distillation step of cognac elaboration. However, in the current study, the formation of these methyl esters was not related to artefacts formed as a result of the temperature used during extraction by SDE, because they were detected with SDE, CLSA, and HS-SPME techniques. Moreover, methyl esters in wines are related to yeast fermentation (Castro et al., 2004). Sparkling wine acetates decrease along ageing time of cava in contact with lees (Riu-Aumatell et al., 2006). The samples used in our study were aged for 14 months. As a result, only isoamyl acetate, hexyl acetate, and 2-fenilethyl were detected. CLSA shows better extraction capacity for acetates than HS-SPME and SDE. However, the main acetate in wine (isoamyl acetate) was detected with all three techniques. Acetal formation was attributed to the reaction between aldehydes and ethanol by temperature or oxidation mechanisms. We found that CLSA and HS-SPME did not detect acetals; therefore we propose that acetals were artefacts formed as a result of the high extraction temperature required when applying SDE. Terpene compounds (lilial and compound numbers 58, 60, 61, 64, and 78) were better determined by SDE than by CLSA or HS-SPME (Fig. 1 and Table 1). Two characteristic norisoprenoids of aged cava, vitispirane isomers and TDN, were identified with all the methods tested. These compounds were markers of ageing because their concentrations increased with ageing (Francioli et al., 1999; Riu-Aumatell et al., 2006). Terpene and norisoprenoid compounds are varietal volatiles released during ageing, and will be better characterized by SDE. Furfurals formed by sugar degradation in the presence of oxygen are usually found in sweet wines that have gone through biological ageing (Schneider et al., 1998). SDE and CLSA showed a greater capacity to determine ethyl 2- furancarboxylate, 2-acetylfuran, and 5-methylfurfural in cava samples than HS-SPME. 4. Conclusion Our findings contribute to a better knowledge of the volatile constituents of cava sparkling wine. Eighty-four compounds were identified using SDE, CLSA, and HS-SPME extraction methods. SDE showed the best extraction rates for all families of volatiles. However, this laborious technique is costly in solvents and time. A good alternative is CLSA because it provides a wide profile of volatile compounds in cava samples. However, CLSA is also lengthy. HS-SPME is a straightforward and fast extraction method which does not require sample treatments. This method allowed us to obtain the most important volatile compounds of each chemical family. Further studies will focus on HS-SPME to assess the occurrence of compounds formed from sugar degradation (furfurals and carbonyl compounds) during biological ageing of sparkling wine. Acknowledgements We are grateful to Segura Viudas SA, Castellblanch SA and Freixenet SA wineries for providing samples. This study was made possible thanks to financial assistance from the Comisión Interministerial de Ciencia y Tecnología (CICYT) (Spain) AGL2005-03451, from Generalitat de Catalunya (Spain), Project 2005SGR-00156 and through a Grant from the Generalitat de Catalunya to the PhD student J. Bosch-Fusté. References Aznar, M., López, R., Cacho, J. F., & Ferreira, V. (2001). Identification and quantification of impact odorants of aged red wines from rioja. GC-Olfactometry, quantitative GC MS, and odor evaluation of HPLC fractions. Journal of Agricultural Food Chemistry, 49, 2924 2929. Bayonove, C., Baumes, R., Crouzet, J., & Günata, Z. (2000). In C Flanzy (Ed.), Enología: Fundamentos Científicos y Tecnológicos. AMV Ediciones (pp. 137 176, Chapter 5). Blaise, A., & Bertrand, A. (2000). In C Flanzy (Ed.), Enología: Fundamentos Científicos y Tecnológicos. AMV Ediciones (pp. 720 742, Chapter 24). Blanch, G. P., Reglero, G., & Herraiz, M. (1996). Rapid extraction of wine aroma compounds using a new simultaneous distillation-solvent extraction device. Food Chemistry, 56(4), 439 444. Bonino, M., Schellino, R., Rizzi, C., Aigotti, R., Delfini, C., & Baiocchi, C. (2003). Aroma compounds of an italian wine (Ruché) by HS-SPME analysis coupled with GC-ITMS. Food Chemistry, 80, 125 133. Câmpeanu, G., Burcea, M., Doneanu, C., Nămolosanu, I., & Visan, L. (1998). GC/MS characterization of the volatiles isolated from the wines obtained from the indigenous cultivar Feteasca Regala. Analysis, 26, 93. Castro, R., Natera, R., Benitez, P., & Barroso, C. G. (2004). Comparative analysis of volatile compounds of fino sherry wine by rotatory and continuous liquid liquid extraction and solid-phase microextraction in conjunction with gas chromatography mass spectrometry. Analytica Chimica Acta, 513, 141 150. Comuzzo, P., Tat, L., Tonizzo, A., & Battistutta, F. (2006). Yeast derivatives (extracts and autolysates) in winemaking: Release of volatile compounds and effects on wine aroma volatility. Food Chemistry, 99(2), 217 230. De la Calle-García, D., Reichenbächer, M., Danzer, K., Hurlbeck, C., Bartzsch, C., & Feller, K. H. (1997). Investigations on wine bouquet components by solid-phase microextraction-capillary gas chromatography (SPME-CGC) using different fibers. Journal of High Resolution Chromatography, 20, 665 668. Díaz-Maroto, M. C., Pérez-Coello, M. S., & Cabezudo, M. D. (2002). Supercritical carbon dioxide extraction of volatiles from spices comparison with simultaneous distillation extraction. Journal of Chromatography A, 947, 23 29. Eggers, N., Kenefick, S., Richardson, S., Wigglesworth, T., & Girad, B. (2003). Evaluation of closed-loop stripping for the isolation of wine aroma compounds from aqueous solution. Americal Journal of Enology and Viticulture, 54, 92 98.

J. Bosch-Fusté et al. / Food Chemistry 105 (2007) 428 435 435 Fernandes, L., Relva, A. M., Gomes da Silva, M. D. R., & Costa Freitas, A. M. (2003). Different multidimensional chromatographic approaches applied to the study of wine malolactic fermentation. Journal of Chromatography A, 995, 161 169. Francioli, S., Guerra, M., Lopez-Tamames, E., Guadayol, J. M., & Caixach, J. (1999). Aroma of sparkling wines by headspace/solid phase microextraction and gas chromatography/mass spectrometry. American Journal of Enology and Viticulture, 50, 404 408. Genovese, A., Dimaggio, R., Lisanti, M. T., Piombino, P., & Moio, L. (2005). Aroma composition of red wines by different extraction methods and gas chromatography SIM/MASS spectrometry analysis. Annali di Chimica, 95, 383 394. Karásek, P., Planeta, J., Varad ová Ostrá, E., Mikešová, M., Goliáš, J., Roth, M., et al. (2003). Direct continuous supercritical fluid extraction as a novel method of wine analysis comparison with conventional indirect extraction and implications for wine variety identification. Journal of Chromatography A, 1002, 13 23. Ledauphin, J., Saint-Clair, J. F., Lablanquie, O., Guichard, H., Fournier, N., Guichard, E., et al. (2005). Identification of trace volatile compounds in freshly distilled calvados and cognac using preparative separations coupled with gas chromatography mass spectrometry. Journal of Agricultural and Food Chemistry, 52, 5124 5134. Liu, M., Zeng, Z., & Tian, Y. (2004). Elimination of matrix effects for headspace solid-phase microextraction of important volatile compounds in red wine using a novel coating. Analytica Chimica Acta, 540, 341 353. López, R., Ferreira, V., Hernández, P., & Cacho, J. (1999). Identification of impact odorants of young red wines made with Merlot, Cabernet Sauvignon and Grenache grape varieties: A comparative study. Journal of the Science of Food and Agriculture, 79, 1461 1467. Malleret, L., Bruchet, A., & Hennion, M. C. (2001). Picogram determination of earthy-musty odorous compounds in water using modified closed loop stripping analysis and large volume injection GC/MS. Analytical Chemistry, 73, 1485 1490. Ortega, C., López, R., Cacho, J., & Ferreira, V. (2001). Fast analysis of important wine volatile compounds development and validation of a new method based on gas chromatographic-flame ionisation detection analysis of dichloromethane microextracts. Journal of Chromatography A, 923, 205 214. Riu-Aumatell, M., Bosch-Fusté, J., López-Tamames, E., & Buxaderas, S. (2006). Development of volatile compounds of cava (Spanish sparkling wine) during long ageing time in contact with lees. Food Chemistry, 95, 237 242. Salinas, M. R., Alonso, G. L., & Esteban-Infantes, J. (1994). Adsorptionthermal desorption-gas chromatography applied to the determination of wine aromas. Journal of Agricultural Food Chemistry, 42, 1328 1331. Sánchez-Palomo, E., Pérez-Coello, M. S., Díaz-Maroto, M. C., González- Viñas, M. A., & Cabezuedo, M. D. (2006). Contribution of free and glycosidically-bound volatile compounds to the aroma of Muscat a petit grains wines and effect of skin contact. Food Chemistry, 95, 279 289. Schneider, R., Baumes, R., Bayonove, C., & Razungles, A. (1998). Volatile compounds involved in the aroma of sweet fortified wines (vins oux naturels) from Grenache noir. Journal Agricultural Food Chemistry, 46, 3230 3237. SciFinder Scholar Ò (version 2006). American Chemical Society. Tat, L., Comuzzo, P., Stolfo, I., & Battistutta, F. (2005). Optimization of wine headspace analysis by solid-phase microextraction capillary gas chromatography with mass spectrometric and flam ionization detection. Food Chemistry, 93, 361 369. Torrens, J., Riu-Aumatell, M., López-Tamames, E., & Buxareras, S. (2004). Volatile compounds of red and white wines by headspace-solidphase microextraction using different fibers. Journal of Chromatographic Science, 42, 110 116. Vannier, A., Brun, O. X., & Feinberg, M. H. (1999). Application of sensory analysis to champagne wine characterisation and discrimination. Food Quality and Preference, 10, 101 107. Vas, G. Y., Gál, J., Harangi, J., Dobó, A., & Vékey, K. (1998). Determination of volatile aroma compounds of Bläufrankisch wines extracted by solid-phase microextraction. Journal of Chromatographic Science, 36, 505 510. Zea, L., Moyano, L., Moreno, J., Cortes, B., & Medina, M. (2001). Discrimination of the aroma fraction of sherry wines obtained by oxidative and biological ageing. Food Chemistry, 75, 79 84.